1
Mapping Macrocystis pyrifera beds from satellite images in New Zealand Sanky Meng, Vincent Zintzen, Helen Curtis and Shane Geange Department of Conservation, National Office, Wellington Summary Recent advances in the resolution of images provided by satellites mean that it is now possible to cost-effectively map the surface canopy of Macrocystis pyrifera beds remotely, which allows for the analysis of time-series data. We present analysis using near infrared bands and normalized difference vegetation index (NDVI) from WorldView-II satellite imagery to map Macrocystis beds in the Marlborough Sounds (Tory Channel and Long Island). The main findings from this study are that: Diver in dense kelp forest at Ulva Island/Te Whawhara Marine Reserve in Stewart Island. 1. Current resolution of satellite images and analytical methods makes possible the large-scale mapping of Macrocystis pyrifera beds in New Zealand waters using this method. 2. It is possible to follow the evolution through time of Macrocystis surface canopy, satellite images being acquired every two weeks. 3. Comparisons of progression or regression of Macrocystis surface canopy between sites can be made. Background Method 1,2,3 Background Method 1. Giant kelp Macrocystis pyrifera (bladder kelp) is a large perennial kelp that forms dense beds with layers of floating surface canopies. 2. These beds are at the base of many temperate coastal food webs, provisioning important habitat. Marlborough Sounds Long Island Tory Channel True colour satellite image False colour satellite image Normalized Difference Vegetation Index Binary result of NDVI Isolation of kelp patches from mainland Final map with kelps in red 1 2 3 4 5 6 4. Although there is some evidence that the extent of some of New Zealand’s Macrocystis beds are declining, there has to date been 3. Because giant kelp is a foundation species that responds to a range of pressures, it is a potentially good indicator species for monitoring the ecological integrity of the New Zealand marine environment. Multispectral satellite images (WorldView-II) were acquired for two areas in the Marlborough Sounds (around Long Island and in Tory Channel). For each location, satellite images were acquired for the winter of 2010 and 2013 when the growth of Macrocystis is maximal 4 . Image resolution: pixel size of 50cm in panchromatic and 2m in multispectral. (NDVI) calculated using red and near infrared bands vegetation Images were geo-referenced in ENVI/ArcMap. 4 images were created: true colour image (Red, Green, Blue), false colour image 1 (Near Infrared, Red, Green), false colour image 2 (Red Edge, Red, Green) and normalized difference vegetation index (NIR- Red/NIR+Red). Vegetation strongly absorbs NIR. The 4 images were compared to identify the canopy of Macrocystis. Band math tool was used to process NDVI into binary result (if NDVI ≥ 0.001 then x = 10 else x = 20). NDVI ≥ 0.001 is characteristic of Results little quantitative evidence to assess the extent and rate of this decline. Tory Channel (26 Sep 2013) Tory Channel (30 Sep 2010) Long Island (30 Sep 2010) Long Island (18 Oct 2013) Images bands: blue (450-510nm), green (510- 580nm), red (630-690nm) and near infrared (NIR, 770-895nm) + 4 additional bands. vegetation. Raster result was converted to vector file. Vector result was edited by cutting and deleting noisy information such as water, coast line and rocks. Macrocystis area: 53,030m 2 Macrocystis area: 45,022m 2 Macrocystis area: 78,265m 2 Macrocystis area: 100,415m 2 In red, surface canopy forming Macrocystis pyrifera, as derived from satellite imagery. The classification results of satellite images could pick up the signal of canopy forming kelps (red layer in above images). Significantly different Macrocystis canopy extent were detected between areas and years. The extent of Macrocystis canopy beds were 53,030m 2 , 45,022m 2 , 78,265m 2 and 100,415m 2 for Long Island 2010, Long Island 2013, Tory Channel 2010, Tory Channel 2013. Prospects Satellite images could provide a powerful tool to remotely map the extent of Macrocystis pyrifera beds. The potential exists for quickly mapping large areas and creating robust reference points for past and future time series. Further tests are required to assess the effect that tidal levels, currents or turbidity might have on the signal picked up by satellites. References 1 Anderson, R. J., Rand, A., Rothman, M. D., Share, A., & Bolton, J. J. (2007). Mapping and quantifying the South African kelp resource. African Journal of Marine Science, 29: 369-378. 2 Cavanaugh, K. C., Siegel, D. A., Kinlan, B. P., & Reed, D. C. (2010). Scaling giant kelp field measurements to regional scales using satellite observations. Marine Ecology Progress Series, 403: 13-27. 3 Cavanaugh, K. C., Siegel, D. A., Reed, D. C., & Dennison, P. E. (2011). Environmental controls of giant-kelp biomass in the Santa Barbara Channel, California. Marine Ecology Progress Series, 429: 1-17. 4 Pirker JG. 2002. Demography, Biomass Production and Effects of Harvesting Giant Kelp Macrocystis Pyrifera (Linnaeus) in Southern New Zealand. PhD thesis thesis, University of Canterbury. 244 pp.

Mapping Macrocystis pyrifera beds from satellite images in ...€¦ · beds with layers of floating surface canopies . 2. These beds are at the base of many temperate coastal food

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Mapping Macrocystis pyrifera beds from satellite images in ...€¦ · beds with layers of floating surface canopies . 2. These beds are at the base of many temperate coastal food

Mapping Macrocystis pyrifera beds from satellite images in New Zealand

Sanky Meng, Vincent Zintzen, Helen Curtis and Shane GeangeDepartment of Conservation, National Office, Wellington

SummaryRecent advances in the resolution of images provided by satellites mean that it is now possible to cost-effectively map the surface canopy of Macrocystis pyrifera beds remotely, which allows for the analysis of time-series data. We present analysis using near infrared bands and normalized difference vegetation index (NDVI) from WorldView-II satellite imagery to map Macrocystis beds in the Marlborough Sounds (Tory Channel and Long Island).The main findings from this study are that:

Diver in dense kelp forest at Ulva Island/Te Whawhara Marine Reserve in Stewart Island.

1. Current resolution of satellite images and analytical methods makes possible the large-scale mapping of Macrocystis pyrifera beds in New Zealand waters using this method.

2. It is possible to follow the evolution through time of Macrocystis surfacecanopy, satellite images being acquired every two weeks.

3. Comparisons of progression or regression of Macrocystis surface canopy between sites can be made.

Background Method 1,2,3Background Method 1. Giant kelp Macrocystis pyrifera

(bladder kelp) is a large perennial kelp that forms dense beds with layers of floating surface canopies.

2. These beds are at the base of many temperate coastal food webs, provisioning important habitat.

MarlboroughSounds

Long Island

Tory Channel

True colour satellite image

False colour satellite image

Normalized Difference

Vegetation Index

Binary result of NDVI

Isolation of kelp patches from

mainland

Final map with kelps in red

1 2 3 4 5 6

4. Although there is some evidence that the extent of some of New Zealand’s Macrocystis beds are declining, there has to date been

3. Because giant kelp is a foundation species that responds to a range of pressures, it is a potentially good indicator species for monitoring the ecological integrity of the New Zealand marine environment.

• Multispectral satellite images (WorldView-II) were acquired for two areas in the Marlborough Sounds (around Long Island and in Tory Channel).

• For each location, satellite images were acquired for the winter of 2010 and 2013 when the growth of Macrocystis is maximal4.

• Image resolution: pixel size of 50cm in panchromatic and 2m in multispectral.

(NDVI) calculated using

red and near infrared bands

vegetation

• Images were geo-referenced in ENVI/ArcMap.• 4 images were created: true colour image (Red, Green, Blue), false colour

image 1 (Near Infrared, Red, Green), false colour image 2 (Red Edge, Red, Green) and normalized difference vegetation index (NIR-Red/NIR+Red). Vegetation strongly absorbs NIR.

• The 4 images were compared to identify the canopy of Macrocystis.• Band math tool was used to process NDVI into binary result (if NDVI ≥

0.001 then x = 10 else x = 20). NDVI ≥ 0.001 is characteristic of

Results

glittle quantitative evidence to assess the extent and rate of this decline.

Tory Channel (26 Sep 2013)Tory Channel (30 Sep 2010)

Long Island (30 Sep 2010) Long Island (18 Oct 2013)

• Images bands: blue (450-510nm), green (510-580nm), red (630-690nm) and near infrared (NIR, 770-895nm) + 4 additional bands.

vegetation.• Raster result was converted to vector file. Vector result was edited by

cutting and deleting noisy information such as water, coast line and rocks.

Macrocystis area:53,030m2

Macrocystis area:45,022m2

Macrocystis area:78,265m2

Macrocystis area:100,415m2

In red, surface canopy forming Macrocystis pyrifera, as derived from satellite imagery.

• The classification results of satellite images could pick up the signal of canopy forming kelps (red layer in above images).

• Significantly different Macrocystis canopy extent were detected between areas and years.

• The extent of Macrocystis canopy beds were 53,030m2, 45,022m2, 78,265m2 and 100,415m2 for Long Island 2010, Long Island 2013, Tory Channel 2010, Tory Channel 2013.

Prospects• Satellite images could provide a powerful tool to remotely map the extent of Macrocystis pyrifera beds.• The potential exists for quickly mapping large areas and creating robust reference points for past and future time series.

• Further tests are required to assess the effect that tidal levels, currents or turbidity might have on the signal picked up by satellites.

References1 Anderson, R. J., Rand, A., Rothman, M. D., Share, A., & Bolton, J. J. (2007). Mapping and quantifying the South African kelp resource. African Journal of Marine Science, 29: 369-378.

2 Cavanaugh, K. C., Siegel, D. A., Kinlan, B. P., & Reed, D. C. (2010). Scaling giant kelp field measurements to regional scales using satellite observations. Marine Ecology Progress Series, 403: 13-27.

3 Cavanaugh, K. C., Siegel, D. A., Reed, D. C., & Dennison, P. E. (2011). Environmental controls of giant-kelp biomass in the Santa Barbara Channel, California. Marine Ecology Progress Series, 429: 1-17.

4 Pirker JG. 2002. Demography, Biomass Production and Effects of Harvesting Giant Kelp Macrocystis Pyrifera (Linnaeus) in Southern New Zealand. PhD thesis thesis, University of Canterbury. 244 pp.